JP4113806B2 - Delivery video storage device and delivery video storage method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a distribution video storage apparatus and a distribution video storage method that extract and store a video of a speaker with an attribute.
[0002]
[Prior art]
Conventionally, video recorded by a video camera is recorded. Then, each member of the parliament, speaker, etc. was cut out and manually input and edited the name of the member of the parliament, questions, the title of the lecture, and the like, and a replayable video was created.
[0003]
In addition, the moderator inputs the text information of the speech content, records the text information linked to the speech information detected from the microphone volume and duration of the conference participant, and records the text information based on the text information. There is a technique for searching for corresponding multimedia information (Patent Document 1).
[0004]
[Patent Document 1]
See, for example, [0010] of JP-A-7-226931 and FIG.
[0005]
[Problems to be solved by the invention]
With the former technique, it takes a lot of labor to manually cut out the corresponding part from the shot video and give the name, title, etc. of the legislators and lecturers to the cut out video, There was a problem that names and titles were given to videos quickly and in large numbers, and it was not possible to search and view them using these names and titles.
[0006]
In the latter technique, the text information input by the presenter is linked to the multimedia information of the section detected from the microphone volume and duration from the participant, and the corresponding multimedia information is searched by searching the text information later. (Video) can be viewed and viewed, but text information can be linked only by extracting and linking video (multimedia information) in a section where the voice from the participant is loud and sustained. However, there is a problem that it is not possible to link appropriate text information for a section where the participant is not introduced or a section where the voice is interrupted or starts.
[0007]
In order to solve these problems, the present invention cuts the section from the start position to the end position of the speaker's speech and captures the attributes of the speaker's name, title, etc. The text information is automatically assigned and stored as a text searchable video so that it can be distributed.
[0008]
[Means for Solving the Problems]
Means for solving the problem will be described with reference to FIG.
In FIG. 1, a camera 1 captures a speaker (such as a member of the Diet) and generates a video.
[0009]
The video cut-out position detection means 3 detects the cut-out start position and end position of the speaker's video from the video.
The video cutout means 4 cuts out the video of the speaker from the video.
[0010]
The headline generation unit 5 generates text information including attributes such as a speaker's name and title to be added to the speaker's video.
The HTML automatic generation means 6 automatically generates a file in which text information consisting of a speaker's video and attributes is described using the HTML language.
[0011]
The upload means 7 uploads an HTML file.
Next, the operation will be described.
The video of the speaker is photographed by the camera 1, the video cutout position detection means 3 detects the cutout start position and end position of the video of the speaker, and the video cutout means 4 is sandwiched between the detected cutout start position and end position. The section video is cut out, and text information including attributes such as the name of the speaker and the title of the voice is added to the video cut out by the header generation unit 5 and the upload unit 7 uploads the video.
[0012]
At this time, the operator uses the input device such as a mouse or a keyboard to specify the start / end positions of the video of the speaker.
Also, let the speaker have a tag with the speaker's ID written on it, and when it is detected that the speaker is within a predetermined range of the speaker base or when a tag is detected, the speaker's video will be cut out, and a certain amount of time has passed. When it is detected that the user has gone out or the tag is no longer detected, it is set as the cutout end position of the video of the speaker.
[0013]
In addition, when a tag that includes the title of the document is attached to a document that describes the content of the speaker, and the speaker places the document with the tag attached to the speaker, or approaches the speaker, The title of the speaker's speech is read from the tag attached to the document and used as an attribute.
[0014]
In addition, text information and video are described in HTML language, uploaded and stored, and video can be searched using text information.
Therefore, for video shot of the speaker in the assembly, etc., by extracting the section from the start position to the end position of the speaker's speech and automatically assigning text information of attributes such as the name and title of the speaker, It can be stored as a text searchable video.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
Next, embodiments and operations of the present invention will be described in detail sequentially with reference to FIGS. In the following explanation, a case will be described as an example of a case where a member is a member of the parliament and a member speaks in the assembly.
[0016]
FIG. 1 shows a system configuration diagram of the present invention.
FIG. 1A shows a conceptual system configuration diagram.
In FIG. 1A, a camera 1 is a video camera that captures a speaker (such as a member of the parliament) and generates a video.
[0017]
The video acquisition means 2 captures video captured by the camera 1.
The video cut-out position detection means 3 detects the cut-out start position and end position of the speaker's video from the video captured by the camera 1 and captured.
[0018]
The video cutout means 4 cuts out a video from the detected cutout start position to the end position.
The headline generation unit 5 generates text information including attributes such as a speaker's name and title (headline) to be added to the video of the speaker.
[0019]
The HTML automatic generation means 6 automatically generates a file (HTML file) in which text information composed of a speaker's video and attributes is described using the HTML language.
[0020]
The upload means 7 uploads an HTML file to the server 8 via the network.
The server 8 accumulates a large number of HTML files, receives search requests from the end user's terminal 9, searches the text information of the HTML files, and downloads (distributes) the corresponding HTML files (also video files). Is.
[0021]
FIG. 1B shows a detailed system configuration diagram. Here, 1 to 9 correspond to the numbers 1 to 9 in FIG. 1A, respectively (the description of 1, 3, 4, 6 to 9 is omitted).
[0022]
In FIG. 1B, the video acquisition unit 2 includes a video generation unit 21, a cut-out source video file 22, and the like.
The video generation means 21 captures a video of a speaker (member) taken by the camera 1.
[0023]
The cut-out source video file 22 is a cut-out source video file (a video file having a compression or resolution, the number of frames, the number of colors, or the like as necessary) captured from the camera 1 by the video generation unit 21.
[0024]
The headline generating means 5 is composed of a headline input means 55 and the like, and takes out a corresponding one from a pre-registered headword dictionary (title dictionary) or puts it in advance on a tag attached to a speaker's document. For example, text information including attributes such as a headword (such as a title) is generated by reading out the written headword (such as a title).
[0025]
Next, the operation of the configuration of FIG. 1 will be described in detail according to the order of the flowchart of FIG.
FIG. 2 is a flowchart for explaining the operation of the present invention (part 1).
[0026]
In FIG. 2, S1 is initialized. Here, the following settings shown in the figure are initialized.
・ Adjustment of camera position and angle ・ Processing to connect to the network ・ Input of initial information such as parliament denomination, meeting name, and annual schedule of the assembly Here, adjustment of the camera position and angle is remarked in the assembly by camera 1 in FIG. Adjust the camera position and angle so that the legislators can be photographed exactly. The network connection processing performs network connection processing (for example, connection to a LAN, connection to a wireless LAN, etc.) for capturing video captured by the camera 1 in a control console (not shown) in real time. The initial input of information such as the denomination of the member of the parliament, the name of the meeting, and the annual schedule of the parliament is pre-stored by entering text information consisting of attributes such as the denomination of the legislator who is going to speak in the parliament, the name of the meeting, and the annual schedule of the parliament. Keep it.
[0027]
In S2, an entry word is input. In this case, a headword (title or the like) of contents to be spoken by a member of the Diet is input (for example, selected and input from accumulated information).
[0028]
In S3, shooting starts. In this case, the shooting start button is pressed, and the camera 1 starts shooting.
In S4, creation of a cut-out source video file is started. In response to the start of shooting by the camera 1 in S3, creation of a cut-out source video file is started.
[0029]
In S5, a member name and a question item are selected. This is done by selecting the member name and question item of the member who is going to speak from the screen.
In S6, the attribute number is temporarily stored. This temporarily stores the attribute number assigned to the text information such as the member name and question item of the member selected in S5.
[0030]
S7 designates a cutout start position. This designates the start position to be cut out from the video shot of the legislator.
S8 designates the cutout end position. This designates the cut-out end position from the video shot of the legislator.
[0031]
In S9, the corresponding section is cut out to a video file. In this method, a video of a member of the Diet taken by the camera 1 is cut out from a cutout start position in S7 to a cutout end position in S8 to create a video file.
[0032]
In S10, the video file name is determined. This confirms the file name to be given to the video file obtained by photographing the legislator cut out in S9 (for example, the photographing date, the identification number representing the meeting, etc. are confirmed as the file name).
[0033]
S11 extracts a headword from the attribute number. This extracts attribute information (conference name, date, member name, member ID, denomination, question item, etc.) from the table (see FIG. 7) storing the attribute information based on the attribute number temporarily stored in S6.
[0034]
In step S12, an HTML file is created. This creates an HTML file (for example, an HTML file in FIG. 8B described later) in which the video file name determined in S10 and the headword (attribute) extracted in S11 are linked.
[0035]
In S13, a video file and an HTML file are uploaded. As a result, the HTML file and the video file are uploaded and stored in the server 8 of FIG.
[0036]
In S14, it is determined whether the photographing is finished. In the case of YES, shooting is ended in S15, and a series of processing is ended. If NO, return to S5 and repeat.
As described above, a situation where a member speaks in the parliament is photographed with the camera 1, a video is cut out during this period, a file name is given, and attribute information of the member (leg member name, meeting name, question item, etc.) is extracted. It is possible to create and link to an HTML file based on the both, upload it to the server 8 quickly and accumulate it, and distribute it.
[0037]
FIG. 3 shows a flowchart (part 2) for explaining the operation of the present invention. FIG. 3 is a flowchart for selecting a headword after designating a cutout start position.
In FIG. 3, S25 designates the cutout start position. This is subsequent to S21 to S24 (same as S1 to S4 in FIG. 2), and the cutout position is designated (the cutout position is designated because the legislator approaches the speech stand).
[0038]
In S26, a member name and a question item are selected. Select the member name and question item of the member who is the speaker.
In S28, the cutout end position is designated.
[0039]
In S29, the corresponding section is cut out to a video file.
In S30, the video file name is determined.
In S31, the attribute number is temporarily stored.
[0040]
In S32, a headword is extracted from the attribute number.
Then, through S32 to S35 (same as S12 to S15 in FIG. 2), an HTML file is created based on the video file and the attributes of the members (members ID, member ID, date, question items, etc.) and linked together. And upload to the server 8.
[0041]
FIG. 4 shows an explanatory diagram (cutout start / end positions) of the present invention. This is a schematic view showing that a hatched section from the cutout start position to the cutout end position is cut out from the video file (cutout source video file) representing the status of the legislator taken by the camera 1 and generated as a cutout video file. Shown in In this case, the video in the section from the cutout start position 8:10 to the cutout end position 8:30 is cut out and used as a cut out video file.
[0042]
As described above, it is possible to automatically cut out a video file (video file after cutout) between the start position and end position by simply specifying the cutout start position and end position from the cutout original video file. It becomes. Then, a file name is given, and attributes (separator name, member ID, question item, etc.) extracted separately can be linked to each other as an HTML file and uploaded to the server 8 as text information.
[0043]
FIG. 5 shows an example screen of the present invention. This screen example is an example of a screen for registering headwords. Here, the screen information is for registering the following information shown in the figure.
・ Meeting name: 1st regular meeting ・ Date: July 1, 2003 ・ Member of the Diet: Taro Fujitsu ・ Question item: Enter or select the above items on the screen for garbage disposal issues as shown in the figure below When the registration button is pressed, registration is made in association with the attribute number as shown in the attribute information table of FIG.
[0044]
FIG. 6 shows an example of the operation screen of the present invention. This is an example of an operation screen of the assembly video distribution system, and includes the following information and various buttons shown in the figure.
・ Meeting name:
·event date:
・ Representative name:
・ Question items:
·Status:
・ Preview screen:
·button:
-Relay start button-Relay end button-Extraction start button-Extraction end button-Edit screen button-Other:
Here, the name of the meeting, the date of the meeting, the name of the member of the parliament, the question item, and the state are the column for selecting the name of the meeting that the member of parliament intends to shoot with the camera 1, the date of the meeting, the name of the member of the parliament speaking, This is a column for selecting and a column for displaying an operation history. The preview screen is an area for displaying a preview video indicated by elapsed time. The relay start button and the relay end button are buttons for starting and ending relay of a member's speech with the camera 1. The cutout start button and the cutout end button are used to manually instruct start and end of clipping from the video of the speech situation of the legislator taken by the camera 1.
[0045]
FIG. 7 is an explanatory diagram of the present invention (example of attribute information table). In the attribute information table, the following information shown in FIG.
・ Attribute number:
・ Meeting name:
·date:
・ Representative name:
・ Representative ID:
・ Sect:
・ Question items:
・ Other:
Here, the attribute number is a unique number assigned to the attribute. The meeting name is the name of the meeting spoken by the member. The date is the date on which the meeting is held. The member name, member ID, and denomination are the member name, member ID, and denomination to speak (question). The question item is a question item asked by a member of the Diet.
[0046]
As described above, by registering and managing the attributes (conference name, date, member name, member ID, question item, etc.) registered in association with the attribute number in the attribute information table, the attribute number is designated and the member It is possible to add the attribute information designated by the attribute number to the video file in the section in which the user said (creates an HTML file and links the two). Then, it is possible to search the attribute information of the text added to the video and view the corresponding video file.
[0047]
FIG. 8 shows an explanatory diagram (HTML) of the present invention.
FIG. 8A shows a screen example, and FIG. 8B shows an HTML file example. The HTML file in FIG. 8B shows an example in which attribute information (see FIG. 7) is linked to a video file. Here, the attribute information shown below is associated with the video file. .
[0048]
・ 1st regular meeting ・ ○○ assembly ・ June 2003 regular meeting question member and question item ・ Representative Taro Fujitsu (ABC party)
・ Waste disposal issues ・ Welfare issues ・ Representative Hanako Fujitsu (Meeting of XYZ)
・ Housing policy Here, when the corresponding video playback button on the right end of the screen shown in FIG. 8A is pressed, the video file linked to the video playback button is automatically played and displayed and sound generated for viewing. It becomes possible.
[0049]
FIG. 9 shows another system configuration diagram of the present invention. Here, 1 to 9 correspond to the numbers 1 to 9 in FIG. 1A, respectively (the description of 1, 2 (21, 22), 3, 4, 6 to 9 is omitted).
[0050]
In FIG. 9, the headline generating means 5 reads a document (question) that the RF tag reader / writer 53 reads the member ID from the RF tag 51 or attaches the RF tag 52 when the member wearing the RF tag 51 approaches the vicinity of the speech board. When the document is placed on the floor, the RF tag reader / writer 53 reads the question item from the RF tag 52, and the headline detection means 54 also displays the legislator ID, headline information, etc. read from the RF tags 51, 52. In this case, the attribute information is automatically generated, and is composed of RF tags 51 and 52, an RF tag reader / writer 53, a headline detecting means 54, and the like.
[0051]
The RF tags 51 and 52 are used for reading data wirelessly. Here, the member ID or attribute information is written in advance and is read as necessary.
[0052]
The RF tag reader / writer 53 reads the member ID from the RF tag 51 attached to the member's jacket or the like when the member approaches the question board (or enters through the gate) and detects the cutout start position or leaves the question table. (Or exits through the gate) detects the cut-out end position, or reads a question item or the like from the RF tag 52 when the question document attached with the RF tag 52 is placed on the floor.
[0053]
The headline detecting means 54 generates attribute information (such as a member of the Diet, a question item) based on the member ID, question item, etc. read by the RF tag reader / writer 53 from the RF tags 51, 52.
[0054]
As described above, when a member of the parliament approaches the speech board (or enters through the gate), the legitimate ID is read from the attached RF tag 51 by the RF tag reader / writer 53 to automatically detect the cutout start position or leave the speech board. (Or go out through the gate) and automatically detect the cut end position, automatically cut out the video between the sections, and automatically generate attribute information (heading items etc.) from the legislator ID, question items, etc. It is possible to automatically create an HTML file, link them together, and automatically upload and store them in the server 8.
[0055]
FIG. 10 is a flowchart for explaining another operation of the present invention. Here, S41 and S49 to S55 are the same as S1 and S9 to S15 in FIG.
[0056]
In FIG. 10, S42 writes a headword to the RF tag. In this case, a member's name (member's ID) is written on the RF tag 51 attached to the member and attached to the member, and a question item is written on the RF tag 52 and attached to the question document.
[0057]
In S43, shooting is started. In this case, the shooting start button is pressed, and the camera 1 starts shooting.
In S44, creation of a cut-out source video file is started. In response to the start of shooting by the camera 1 in S43, creation of a cut-out source video file is started.
[0058]
In S <b> 45, the name of the legislator is detected from the RF tag 51. This is detected by the RF tag reader / writer 53 reading the member ID from the RF tag 51 when the member approaches the floor (or enters through the gate).
[0059]
In S46, the question item is detected from the RF tag 52, and the attribute number is temporarily stored. When a member of the Diet puts the question document attached to the RF tag 52 on the floor, the RF reader / writer 53 reads the question item from the RF tag 52 and temporarily stores the attribute number of the read question item.
[0060]
In S47, a cutout start position is designated.
In S48, the cutout end position is designated.
Then, an HTML file is automatically created on the basis of the video of the section from the cutout start position to the cutout end position and the attribute information such as the question item by the procedure of S49 to S55 (S9 to S15 in FIG. 2 described above). The server 8 can be uploaded and stored.
[0061]
As described above, the situation in which the member of the assembly speaks in the assembly is photographed with the camera 1, the video between this is cut out, the file name is given, and the member's attribute information automatically read by the RF tags 51 and 52 (the member's name, meeting name, question) It is possible to automatically create and link to an HTML file based on the item etc., upload it to the server 8 and store it for distribution.
[0062]
FIG. 11 is a flowchart for explaining another operation of the present invention. Here, S61 to S64 and S69 to S75 correspond to S41 to S44 and S49 to S55 of FIG.
[0063]
In FIG. 11, S65 detects the name of the legislator from the RF tag 51 (when the legitimate unit is cut out, the cut-out start position is detected). This is detected by the RF tag reader / writer 53 reading the member ID from the RF tag 51 when the member approaches the floor (or enters through the gate). At this time, when the member is cut out for each member (each member), it is detected as the cut-out start position every time the member ID is read from the RF tag 51.
[0064]
In S66, the question item is detected from the RF tag 52, and the attribute number is temporarily stored. This is because when a member of the Diet puts the question document attached to the RF tag 52 on the floor (or turns the page), the RF reader / writer 53 reads the question item from the RF tag 52, and the attribute of the read question item Temporarily store the number. In this case, when the question item unit (for each question item) is cut out, it is detected as the cut-out start position every time the question item is read from the RF tag 52 (in this case, one question item is described for each page of the question document, It is assumed that an RF tag 52 is attached to each page).
[0065]
In S67, when the question sheet is removed from the predetermined position, when it is cut out in units of question items, it is detected as a cut-out end position.
In S68, when cutting out in units of members, when the member passes the gate for the second time, it is detected as the cut-out end position. In the case of cutting out in units of members, when the member wearing the RF tag 51 passes the gate for the second time (or goes out of the gate), the end position is detected.
[0066]
According to the procedure from S69 to S75 (S49 to S55 in FIG. 10 described above), the video of the segment from the segmentation start position to the segmentation end position (or the segment from the segmentation start position to the next segmentation start position), and attributes such as question items An HTML file can be automatically created based on the information and uploaded to the server 8 for storage.
[0067]
As described above, a situation where a member of the Diet speaks in the assembly is photographed with the camera 1, and a video is automatically cut out based on a member ID automatically read from the RF tag 51 to automatically generate a video file and read from the RF tag 52. It is possible to automatically create and link to an HTML file based on attribute information (a member of the Diet, a meeting name, a question item, etc.) and link it to the server 8 for easy and quick storage and distribution.
[0068]
FIG. 12 is an explanatory diagram of the present invention (when only the cutout start position is detected). This is because when a member of the parliament wearing the RF tag 51 approaches the speech stand (or when passing the gate and approaches the speech stand), or when a questionnaire attached with the RF tag 52 is placed on the speech stand (or (When turning the page of the question document), when the data (member ID, question item, etc.) is read from these RF tags 51 and 52, it is detected as a video cutout start position, and video cutout is started from the cutout original video. FIG. 6 schematically shows a state where the next video cutout start position is set as a video cutout end position and cutout is completed. For example, data (a member of the Diet ID or a question item) is detected from either of the RF tags 51 and 52 after 0 to 10 minutes from the base point at the left end, and is set as a clipping start position, and from the RF tag 51 or 52 after the next 20 minutes. When data (a member of the Diet or a question item) is detected, it is detected as a cutout end position, and is set as a cutout start position of the next video.
[0069]
【The invention's effect】
As described above, according to the present invention, for a video of a speaker or the like taken in a parliament or the like, the section from the start position to the end position of the speaker's speech is cut out and the attributes of the speaker's name, title, etc. Since the text information is automatically assigned, the video of the speaker (such as a member of the parliament) is automatically cut out, and the text information consisting of attributes such as title and name is linked and stored easily and quickly. It becomes possible. As a result, it is possible to specify and search the attribute information (name, title, etc.) of the text linked and stored with the video, and to easily view (distribute) the corresponding video.
[Brief description of the drawings]
FIG. 1 is a system configuration diagram of the present invention.
FIG. 2 is a flowchart (part 1) illustrating the operation of the present invention.
FIG. 3 is a flowchart (part 2) illustrating the operation of the present invention.
FIG. 4 is an explanatory diagram (cutout start / end positions) of the present invention.
FIG. 5 is a screen example (entry word registration) of the present invention.
FIG. 6 is an example of an operation screen according to the present invention.
FIG. 7 is an example of an attribute information table of the present invention.
FIG. 8 is an explanatory diagram (HTML) of the present invention.
FIG. 9 is another system configuration diagram of the present invention.
FIG. 10 is a flowchart illustrating another operation of the present invention.
FIG. 11 is a flowchart illustrating another operation of the present invention.
FIG. 12 is an explanatory diagram of the present invention (when only the cutout start position is detected).
[Explanation of symbols]
1: Camera 2: Video acquisition means 21: Video generation means 22: Cutout source video file 3: Video cutout position detection means 4: Video cutout means 5: Headline generation means 51, 52: RF tag 53: RF tag reader / writer 54: Headline Detection means 55: Headline input means 6: HTML automatic generation means 7: Upload means 8: Server 9: End user (terminal)

Claims (2)

For video taken with a camera, in a distribution video storage device that cuts out the video of the speaker and assigns it to the video,
A camera that captures the video of the speaker,
When the speaker has the first wireless tag in which the ID of the speaker is written and the speaker is detected to be within a predetermined range of the speaker, or when the first wireless tag is detected , the camera The cutout start position of the speaker's video from the captured video, and the cutout end position of the speaker's video when it is detected that the speaker has gone out of the range after a certain period of time or the first wireless tag is no longer detected first detecting means for detecting as,
Attach a second wireless tag with at least the title of the document to the document describing the content of the statement, and the speaker puts the document attached with the second wireless tag on the statement table or approaches the statement table A second detection means that at least reads the title of the message from the second wireless tag attached to the document and sets the attribute;
The video includes an attribute including the name of the speaker and the title of the speech detected by the second detection means by cutting out the video of the section sandwiched between the extraction start position and the end position detected by the first detection means. A distribution video storage apparatus comprising means for adding text information. For the video taken by the camera, in the delivery video storage method of cutting out the video of the speaker and adding and storing it,
It has a camera that takes pictures of the speakers,
When the speaker has the first wireless tag in which the ID of the speaker is written and the speaker is detected to be within a predetermined range of the speaker, or when the first wireless tag is detected , the camera The cutout start position of the speaker's video from the captured video, and the cutout end position of the speaker's video when it is detected that the speaker has gone out of the range after a certain period of time or when the first wireless tag is no longer detected A first detection step for detecting as
Attach a second wireless tag with at least the title of the document to the document that describes the content of the speaker's speech, and the speaker places the document with the second wireless tag on the floor. A second detection step that reads at least the title of the speaker's speech from the second wireless tag attached to the document when it approaches the table,
The video includes an attribute including the name of the speaker and the title of the speech detected by the second detection step by cutting the video of the section sandwiched between the clipping start position and the end position detected by the first detection step. A distribution video storage method including the step of adding text information.

Images